From: Tom Lane <tgl@sss.pgh.pa.us>
To: Matt Gibbins <matt_gibbins@fastmail.com.au>
cc: pgsql-admin@lists.postgresql.org
Subject: Re: Segmentation fault when running queries in sequence
In-reply-to: <94eaab7c-b764-418e-a623-c6b4911f4e32@fastmail.com.au>
References: <728d1797-efdd-4298-8ab2-042a28ddaae2@fastmail.com.au>
 <1282819.1695479196@sss.pgh.pa.us>
 <a1ca0600-4dd5-49c7-aad0-24b6b4fcd410@fastmail.com.au>
 <1418006.1695513406@sss.pgh.pa.us>
 <94eaab7c-b764-418e-a623-c6b4911f4e32@fastmail.com.au>
Comments: In-reply-to Matt Gibbins <matt_gibbins@fastmail.com.au>
	message dated "Sun, 24 Sep 2023 10:46:25 +1000"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1430982.1695516874.1@sss.pgh.pa.us>
Date: Sat, 23 Sep 2023 20:54:34 -0400
Message-ID: <1430983.1695516874@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/1430983.1695516874%40sss.pgh.pa.us>
Precedence: bulk

Matt Gibbins <matt_gibbins@fastmail.com.au> writes:
> I executed continue and the gdb reported '[Detaching after fork from 
> child process 2790431]' for several sequences and the remote process 
> then crashed.

I think this means that you attached to the parent postmaster
process, not to the per-session child process where the crash
will occur.

On the whole, attaching to a live server process is kind of the
hard way to do this anyhow.  I'd recommend enabling saving of
core dump files and then gdb'ing the core file.  That avoids
needing to identify in advance the process that is going to crash.

			regards, tom lane